Considering a PhD in CS? Work in industry first!



I’ve been thinking about my graduate school experience for a while now and I feel like I’ve come to a very different conclusion from the advice I typically got (or at least what I listened to) when I considered graduate school. This is my second contribution to the genre of grad school advice (the other being aimed at MS students; this one is primarily aimed at people considering a PhD). I’m probably going to fail to convince most of my readers in the same way that most advice I read didn’t really resonate with me (the problem with advice is that the giver is drawing from their experiences, and the receiver has no understanding of the giver’s frame of reference.) This is certainly how I know feel about most of the advice I read when I started — the authors certainly meant well, and even said the right things, but the wide chasm in experience between us meant that I didn’t really understand what they meant.

The point I’m trying to make can be nicely summarized as “Hey Undergrad, Go work; get a PhD later." I strongly believe most undergraduates considering a PhD will be better served getting a job at the end of their Bachelor’s degree (the argument can also easily be extended to Master’s grads as well). IMO, going to graduate school to get a PhD right out of a BS or MS program is only suitable for a vanishingly small number of people. Most other advice on this topic tends to be the exact opposite (this advice is typically aimed at the small minority of students who are ready for the PhD at this age, but IMHO is terrible advice for the general student). Better advice columns say something like “understand why you want a PhD and if you’re ready for it”. This is germane advice that succumbs to the trap I discussed earlier (of being almost obvious to the person doling it out, and completely meaningless to the person receiving it.)

Here’s a bit of personal context, in the hope that you’ll understand where I’m coming from: I was a young 23 year old when I decided to switch to the PhD program at Stony Brook University. I was nearing the end of my MS degree, had spent a grand total of 3 months in an industry internship (which went fine), and was having a grand old time working on a research project that would eventually go nowhere (I was not ready for it, but more on this later). My plan at the time was to graduate, get a job, work for a couple of years, and come back to grad school after a couple of years. I should have stuck to this plan, but not for the reasons I had then. I attended an information session about converting to the PhD program primarily because I was interested in coming back after a couple of years. My research adviser at the time was also at the information session. We had an extended conversation after the session, and he argued that I should consider doing the PhD immediately, primarily because the numbers argued that I wouldn’t come back (I’m not sure if there is a thoroughly assessed number, but I’ve heard a lot of professors say that >90% of students who say they will return for a PhD never do). Long story short, I applied to the PhD program essentially overnight (the application deadline for the Spring semester was in a week), bid goodbye to my graduating MS friends, and dove into the PhD, and here I am 6 years later.

Looking back, I wish I had stuck to my original plan, but not for the reasons I had back then. I should have graduated that December, and gotten a job as a software engineer. I’m fairly certain I would have come back to get a PhD after working for a few years. Would I have come back for sure? Hard to say. It’s also another entirely different matter about whether I would have gotten into a PhD program after a few years working on some random widget in industry (the most likely case for most fresh out of school engineers). I can’t be sure of it of course, but from self-reflection I believe that I would have returned: I was very set on getting a PhD, and I was in awe of a lot of people at the company I interned at who seemed to be wizards and the common denominator was that they all had PhDs. I had a strong enough relationship with several professors at Stony Brook University, that I would probably have been admitted if I’d applied a few years later. I could just well be deluded. Further, a real paycheck money is a very hard thing to give up. So, who knows 🤷🏽‍♂️. That said, if I hadn’t come back for a PhD, that wouldn’t necessarily have been a bad thing either, as I’ll argue later in this article.

Allow me to explain why I believe most undergraduates considering a PhD should go work, and get a PhD later:

  1. Most students have education loans: Don’t go to grad school if you have college loans. Compound interest is not something to trifle with. You should instead make it your friend by working for a couple of years, clearing said loans (should be possible to be rid of them quickly enough given how much CS salaries currently are) and saving some money. That way when something unplanned happens in grad school, like your car breaking down, you won’t really be in a pickle. Other things you could do if you had some money saved up: depressed/burned out and want to take a vacation? No problem! Dip into your savings and go recharge! Family emergency and you need to buy an expensive flight ticket? No need to pile it on a credit card and worry about how you’re going to pay it off.
  2. If you don’t have any education loans, make compound interest work for you while you are in graduate school! Squirrel away some money in those years before you return for graduate school, and watch it grow as you earn peanuts for 5 to 6 years in grad school. If the market is anything like it has been for the last decade during your stint in graduate school, your money would likely have tripled or sextupled if you’d invested everything in Amazon for example (which is a stupidly risky move that I’m not advocating for…).
  3. Most students have no understanding of the opportunity cost.
  • Financial opportunity cost. Most of my friends were starting out in entry-level jobs and so the financial opportunity cost didn’t seem to be much when I made the decision to join graduate school. However, in the 6 years I spent making about 50k per annum (on average), my friends got promoted and made tons more money. My opportunity cost (from talking to my friends who left with an MS) has worked out to be somewhere around $600,000 over the past 6 years, or 100k per year on average. Life isn’t all about money, for sure, but being realistic about the financial opportunity cost is very important. I’m happy with the outcome of the trade I made, overall: I’ll be more than fine financially after I graduate (I am lucky to be in a high paying profession as a computer scientist), but I think the decision to pay the opportunity cost should be an informed one.
  • Temporal opportunity cost. The path to a PhD requires something that I was in short supply of when I started: maturity. I’m defining maturity broadly to encompass everything from understanding the value of my and others’ time, knowing myself (what makes me tick, how I like to work, the unique traits I bring to project, my strengths and weaknesses, etc.), to understanding your standing in the world.
    • I started the PhD as a 23 year old who didn’t understand the value of time. I didn’t value my own time, nor did I value others’. I didn’t understand the investment that society was making in me by allowing me to be a graduate student (effectively removing myself from the productive labor force). I understood that the taxpayer was subsidizing me because they saw it as an investment in the future, but I didn’t understand what was required of me. Maybe more importantly, I didn’t understand the value of my own time, of the time that I was investing into this journey. I didn’t think to maximize my investment. I’ve had an amazing time in grad school; don’t get me wrong. I just think the outcome may have been the same whether I came to get a PhD straight out of my MS, or after working for a few years.
    • Understanding yourself: Knowing you’re strengths and weaknesses is a very important part of making the most of your PhD. I didn’t know how I liked to work, what I liked to work on, what kinds of roles I’m most drawn to (PhDs in CS are team sports; less team sport-ey than industry from what I hear though), how to manage my time, how to hold myself accountable, how to build discipline, etc. I spent a lot of my time during the PhD learning all of this. My weaknesses (which I’m still working on) led to an unnecessarily extended PhD (IMO at least). Had I spent a few years in industry (making 3x the PhD pay; see point 2 above about compound interest), I would have gone through the same self-discovery process of the early 20s (I like to call it the party-all-night-and-philosophize-all-day phase of life) and then come back to get a PhD once I was more mature. This would have had many upsides: less stress through the experience, better mental health overall (see point below), and better financial health as well.
      • I think most graduate students come in with the expectation that they are outliers. This is true for the most part, the problem is the population they are considering. Most PhD students are outliers in the general populace (in the few narrow dimensions a PhD program measures for), but what that means is everyone around you in grad school is an outlier, as well (this isn’t a monopoly of academia, of course). This can lead to very volatile social situations: most people are either in denial (everyone thinks they are the best; on average, its safe to assume you are not the best anywhere), or anxiety-ridden (imposter syndrome, feelings of inadequacy, etc.), or in so many cases, both.

In short, I think most people considering a PhD straight out of undergrad are better off taking a few years to understand themselves in a cushy environment (that we Computer Scientists are privileged to enjoy). You go to industry and you spend a couple of years growing into your role and as a person. Two paths seem to be viable after this:

  1. You feel a little intellectually stifled and want to go back to a PhD program. You are more confident of yourself and reach out to others to understand what a PhD really is, and think hard about what you want from it and why. Having understood all of that you identify what you want to work on, who you want to work with, etc., set realistic expectations and go into this journey with a plan, and knock it out of the park (with the attendant challenges and setbacks and growth and all of that).
  2. You find that you’re thriving in industry and while life in general isn’t all rosey (it never is), it’s great as it is! You decide not to get a PhD and go on to do great things anyways. You found a meaningful place to contribute to society and that’s what matters. A degree doesn’t have monopoly on that.